Searching the tender notices that publish every day in open tendering websites is a common way for finding business opportunity in public procurement. The heterogeneity of tender notices from various tendering marketplaces is a challenge for exploiting semantic technologies in the tender search.
Most of the semantic matching approaches require the data to be structured and integrated according to a data model. But the integration process can be expensive and time-consuming especially for multi-source data integration.
In this paper, a product search mechanism that had been developed in an e-procurement platform for matching product e-catalogues is applied to the tender search problem. The search performance has been compared using two procurement vocabularies on searching tender notices from two major tender resources.
The test results show that the matching mechanism is able to find tender notices from heterogeneous resources and different classification systems without transforming the tenders to a uniform data model.
